[Dbix-class] Restricting Result Sets

[Matt S Trout]

On Sat, Nov 08, 2008 at 10:31:52AM -0800, Ovid wrote:
> --- On Fri, 7/11/08, Eden Cardim <edencardim at gmail.com> wrote:
> 
> > <publiustemp-dbic at yahoo.com> wrote:
> > > Because we use the resultset methods heavily,
> > particularly the pager.  We'd have to find a minimal
> > subset of behavior we'd be willing to support and wrap
> > that in our our resultset wrapper.  Yuck.
> > 
> > I think you need DBIx::Class::Schema::RestrictWithObject
> 
> We already looked at that.  As far as we can tell, it's just syntactic sugar for extra "where" conditions.  It doesn't satisfy our requirement because we need to filter on conditions which cannot be expressed via SQL.

Then I don't understand how the pager is any use - it really wants you to
be able to COUNT()

There's no reason you can't do $rs->set_cache([ <arbitrary perl> ]) though.

The other approach would be to create a custom class/object with an
inflate_result method that skips the nonexistant objects - the return value
of inflate_result is actually in list context and DBIC will keep fetching
rows off the DB until it gives it a row back (and correctly stash >1 row
object returned).

I'd be happy to help further with either of these for payment in the form of
a cookbook entry :)

-- 
      Matt S Trout       Need help with your Catalyst or DBIx::Class project?
   Technical Director                    http://www.shadowcat.co.uk/catalyst/
 Shadowcat Systems Ltd.  Want a managed development or deployment platform?
http://chainsawblues.vox.com/            http://www.shadowcat.co.uk/servers/